Night thoughts on resilience and shocks

I've been putting too much trust in distributed networks.  That's where I ended up last night, after driving through the east coast earthquake's aftereffects.

USGS map

We were in Virginia when the quake hit.  Nobody and nothing got hurt by the shock, although we were surprised (first quake for all but one of us in the house).  Naturally we turned to the internet for information and news - avoiding tv - and were richly rewarded.  Social media offered first, freshest stuff.

Networks: power remained on.  Internet stayed up.  Our computers worked.  Beyond that, the outside world of networks thrummed: ISPs, users, etc.

Cheered, I got ready to drive to a nearby meeting.  Normally I'm a bit obsessive about travel planning, especially when it involves driving: Google Mapping my route (checking several alternates and StreetView), printing the results, adding some local maps (which I annotate and save in carefully organized folders).  When I drive, I monitor distances via odometer.  And yes, I violate gender norms by asking strangers for directions.

This time I decided to rely on Onstar.  If you haven't used the thing, it's a phone + GPS-like system built into your car.  You press a button to call them, and they come up with a Mapquest-like route.  OnStar then calls your back, and talks you through the trip. Usually it's pretty handy.

But yesterday the thing simply wouldn't respond.  I sat in the parking lot, waiting patiently (at first) as call after call timed out.  Then I started driving, working from memory of Google Maps. I hadn't printed anything, because I trusted OnStar.  The map I had didn't have the right local details (state level).  As the car worked its way down a road, my memory got fuzzy, and I couldn't recall the right turnoff.  OnStar kept failing.

Now lost, and sinking into horrendous traffic, I pulled out my cell phone to call OnStar's 1-800 number.  Things got weirder - it took a while to get to an operator, who then put me on hold for more time.  Next operator's sound quality quickly degraded to static.  I called again, working the phone tree from start, enjoying another delay as traffic stalled around me and my meeting time approached.  A tired-sounding OnStar operator came on line to agree that yes, there are technical issues, but no, she couldn't explain them.

I began to worry.  I was lost in a town I didn't know, and my network support was failing.  So I pulled over and tried to contact Ceredwyn for help over the cell phone.  Four calls simply failed.  Texting was iffy.

At this point comprehension started seeping into my driving-addled brain.  The earthquake hadn't done physical damage, but the various networks were probably being overwhelmed by usage spikes.  An hour earlier I'd retweeted a call from New York's police department to get people off of emergency lines.  Our social reaction to the quake was stressing our communications apparatus.

Eventually my wife coaxed OnStar back to life, and OnStar managed to reach me in the car.  I drove around, found the meeting, humbly apologized, and got back to work.

So what to learn from this?  First, our American networks are pretty fragile, if a non-lethal, basically non-destructive event can sap them.  Note that weather was fine (no hurricane yet), and this was a very advanced part of American infrastructure (northeastern Virginia) (one nuclear plant).  It's a cliche to observe that modern civilization is fairly thin on top, but important to feel it through lived experience.

Second, I'm too reliant on said networks, at least in a live, synchronous way.  Obviously I should have printed stuff out to carry, and committed more details to memory.  An attitude from circa 2005, rather than 2011.  More, a forward-looking attitude - assuming nothing gets better, but probably worse - means it's essential to prepare for failure in travel.